@ARTICLE{Giuseppi2021ijc,
  doi = {10.1080/00207179.2021.1913516},
  url = {https://doi.org/10.1080/00207179.2021.1913516},
  year = {2021},
  month = apr,
  publisher = {Informa {UK} Limited},
  pages = {1--12},
  author = {Alessandro Giuseppi and Antonio Pietrabissa},
  title = {Bellman's principle of optimality and deep reinforcement learning for time-varying tasks},
  journal = {International Journal of Control},
  document_type={Article},
  abstract = { This paper presents the first framework (up to the authors' knowledge) to address time-varying objectives in finite-horizon Deep Reinforcement Learning (DeepRL), based on a switching control solution developed on the ground of Bellman's principle of optimality. By augmenting the state space of the system with information on its visit time, the DeepRL agent is able to solve problems in which its task dynamically changes within the same episode. To address the scalability problems caused by the state space augmentation, we propose a procedure to partition the episode length to define separate sub-problems that are then solved by specialised DeepRL agents. Contrary to standard solutions, with the proposed approach the DeepRL agents correctly estimate the value function at each time-step and are hence able to solve time-varying tasks. Numerical simulations validate the approach in a classic RL environment. }
}

This paper presents the first framework (up to the authors' knowledge) to address time-varying objectives in finite-horizon Deep Reinforcement Learning (DeepRL), based on a switching control solution developed on the ground of Bellman's principle of optimality. By augmenting the state space of the system with information on its visit time, the DeepRL agent is able to solve problems in which its task dynamically changes within the same episode. To address the scalability problems caused by the state space augmentation, we propose a procedure to partition the episode length to define separate sub-problems that are then solved by specialised DeepRL agents. Contrary to standard solutions, with the proposed approach the DeepRL agents correctly estimate the value function at each time-step and are hence able to solve time-varying tasks. Numerical simulations validate the approach in a classic RL environment.

@ARTICLE{Valensise21,
author = {Carlo M. Valensise and Alessandro Giuseppi and Giulio Cerullo and Dario Polli},
journal = {Optica},
keywords = {Beam splitters; Neural networks; Nonlinear spectroscopy; Optical amplifiers; Parametric down conversion; White light},
number = {2},
pages = {239--242},
publisher = {OSA},
title = {Deep reinforcement learning control of white-light continuum generation},
volume = {8},
month = {Feb},
year = {2021},
document_type={Article},
url = {http://www.osapublishing.org/optica/abstract.cfm?URI=optica-8-2-239},
doi = {10.1364/OPTICA.414634},
abstract = {White-light continuum (WLC) generation in bulk media finds numerous applications in ultrafast optics and spectroscopy. Due to the complexity of the underlying spatiotemporal dynamics, WLC optimization typically follows empirical procedures. Deep reinforcement learning (RL) is a branch of machine learning dealing with the control of automated systems using deep neural networks. In this Letter, we demonstrate the capability of a deep RL agent to generate a long-term-stable WLC from a bulk medium without any previous knowledge of the system dynamics or functioning. This work demonstrates that RL can be exploited effectively to control complex nonlinear optical experiments.},
}

White-light continuum (WLC) generation in bulk media finds numerous applications in ultrafast optics and spectroscopy. Due to the complexity of the underlying spatiotemporal dynamics, WLC optimization typically follows empirical procedures. Deep reinforcement learning (RL) is a branch of machine learning dealing with the control of automated systems using deep neural networks. In this Letter, we demonstrate the capability of a deep RL agent to generate a long-term-stable WLC from a bulk medium without any previous knowledge of the system dynamics or functioning. This work demonstrates that RL can be exploited effectively to control complex nonlinear optical experiments.

@misc{patent,
  title={Brevetto:it-102018000002114, {APP-CI}, Assistente predittivo personalizzato per il controllo di interfacce uomo-computer per pazienti con disabilità motorie basato su metodi di Model Predictive Control e Machine Learning},
  author={Febo Cincotti  and Alessandro Giuseppi  and Antonio Pietrabissa  and Lorenzo {Ricciardi Celsi}  and Cecilia Poli  and D.G. Ferriero },
  year={2020},
  document_type={Patent},
}

@ARTICLE{Priscoli20201,
author={Priscoli, F.D. and Giuseppi, A. and Lisi, F.},
title={Automatic transportation mode recognition on smartphone data based on deep neural networks},
journal={Sensors (Switzerland)},
year={2020},
volume={20},
number={24},
pages={1-16},
doi={10.3390/s20247228},
art_number={7228},
abstract={In the last few years, with the exponential diffusion of smartphones, services for turn-by-turn navigation have seen a surge in popularity. Current solutions available in the market allow the user to select via an interface the desired transportation mode, for which an optimal route is then computed. Automatically recognizing the transportation system that the user is travelling by allows to dynamically control, and consequently update, the route proposed to the user. Such a dynamic approach is an enabling technology for multi-modal transportation planners, in which the optimal path and its associated transportation solutions are updated in real-time based on data coming from (i) distributed sensors (e.g., smart traffic lights, road congestion sensors, etc.); (ii) service providers (e.g., car-sharing availability, bus waiting time, etc.); and (iii) the user’s own device, in compliance with the development of smart cities envisaged by the 5G architecture. In this paper, we present a series of Machine Learning approaches for real-time Transportation Mode Recognition and we report their performance difference in our field tests. Several Machine Learning-based classifiers, including Deep Neural Networks, built on both statistical feature extraction and raw data analysis are presented and compared in this paper; the result analysis also highlights which features are proven to be the most informative ones for the classification. © 2020 by the authors. Licensee MDPI, Basel, Switzerland.},
author_keywords={Artificial neural networks;  Machine learning;  Transportation model recognition},
keywords={5G mobile communication systems;  Deep learning;  Deep neural networks;  Learning systems;  Multimodal transportation;  Smartphones;  Transportation routes, Distributed sensor;  Enabling technologies;  Machine learning approaches;  Real-time transportation;  Statistical feature extractions;  Transportation mode;  Transportation planners;  Transportation system, Neural networks},
document_type={Article},
}

In the last few years, with the exponential diffusion of smartphones, services for turn-by-turn navigation have seen a surge in popularity. Current solutions available in the market allow the user to select via an interface the desired transportation mode, for which an optimal route is then computed. Automatically recognizing the transportation system that the user is travelling by allows to dynamically control, and consequently update, the route proposed to the user. Such a dynamic approach is an enabling technology for multi-modal transportation planners, in which the optimal path and its associated transportation solutions are updated in real-time based on data coming from (i) distributed sensors (e.g., smart traffic lights, road congestion sensors, etc.); (ii) service providers (e.g., car-sharing availability, bus waiting time, etc.); and (iii) the user’s own device, in compliance with the development of smart cities envisaged by the 5G architecture. In this paper, we present a series of Machine Learning approaches for real-time Transportation Mode Recognition and we report their performance difference in our field tests. Several Machine Learning-based classifiers, including Deep Neural Networks, built on both statistical feature extraction and raw data analysis are presented and compared in this paper; the result analysis also highlights which features are proven to be the most informative ones for the classification. © 2020 by the authors. Licensee MDPI, Basel, Switzerland.

@CONFERENCE{Ornatelli20200467,
author={Ornatelli, A. and Tortorelli, A. and Giuseppi, A.},
title={Iterative MPC for Energy Management and Load Balancing in 5G Heterogeneous Networks},
journal={2020 11th IEEE Annual Ubiquitous Computing, Electronics and Mobile Communication Conference, UEMCON 2020},
year={2020},
pages={0467-0471},
doi={10.1109/UEMCON51285.2020.9298113},
art_number={9298113},
abstract={Multi-Access Heterogeneous Networks introduced a step forward in modern communication networks allowing the provision of reliable and efficient broadband services. However, heterogeneous networks imply a burden of complexity in the integration, coordination and QoS management processes thus complicating the satisfaction of users' requirements. The aim of the present work is to address the above-mentioned issues by developing a mathematical framework for optimizing resource usage in 5G heterogeneous networks. More in detail, the optimization will take into account both the network's load and energy consumption simultaneously. The proposed approach, based on Model Predictive Control, will be compared with other control strategies for validation and performance comparison. © 2020 IEEE.},
author_keywords={Energy Management;  Heterogenous Networks;  Load Balancing;  Model Predictive Control;  Multi-Access Networks},
keywords={Energy utilization;  Heterogeneous networks;  Mobile telecommunication systems;  Model predictive control;  Predictive control systems;  Ubiquitous computing, Broadband service;  Control strategies;  Mathematical frameworks;  Multiaccess;  Performance comparison;  QOS management;  Resource usage, 5G mobile communication systems},
document_type={Conference Paper},
}

Multi-Access Heterogeneous Networks introduced a step forward in modern communication networks allowing the provision of reliable and efficient broadband services. However, heterogeneous networks imply a burden of complexity in the integration, coordination and QoS management processes thus complicating the satisfaction of users' requirements. The aim of the present work is to address the above-mentioned issues by developing a mathematical framework for optimizing resource usage in 5G heterogeneous networks. More in detail, the optimization will take into account both the network's load and energy consumption simultaneously. The proposed approach, based on Model Predictive Control, will be compared with other control strategies for validation and performance comparison. © 2020 IEEE.

@CONFERENCE{Giuseppi2020594,
author={Giuseppi, A. and Maaz Shahid, S. and De Santis, E. and Ho Won, S. and Kwon, S. and Choi, T.},
title={Design and Simulation of the Multi-RAT Load-balancing Algorithms for 5G-ALLSTAR Systems},
journal={International Conference on ICT Convergence},
year={2020},
volume={2020-October},
pages={594-596},
doi={10.1109/ICTC49870.2020.9289485},
art_number={9289485},
abstract={This paper introduces algorithms for the multi-RAT load balancing function to maximize QoE in terrestrial and satellite combined system developed in the 5G-ALLSTAR project. The pros and cons of the considered algorithms are described and the simulator is also described in the paper with the on-going performance evaluation processes. © 2020 IEEE.},
author_keywords={5G terrestrial and satellite;  multi-connectivity;  multi-RAT load balance},
keywords={Combined system;  Design and simulation;  Load balancing algorithms, Rats},
document_type={Conference Paper},
}

This paper introduces algorithms for the multi-RAT load balancing function to maximize QoE in terrestrial and satellite combined system developed in the 5G-ALLSTAR project. The pros and cons of the considered algorithms are described and the simulator is also described in the paper with the on-going performance evaluation processes. © 2020 IEEE.

@ARTICLE{Kim2020669,
author={Kim, J. and Casati, G. and Cassiau, N. and Pietrabissa, A. and Giuseppi, A. and Yan, D. and Calvanese Strinati, E. and Thary, M. and He, D. and Guan, K. and Chung, H. and Kim, I.},
title={Design of cellular, satellite, and integrated systems for 5G and beyond},
journal={ETRI Journal},
year={2020},
volume={42},
number={5},
pages={669-688},
doi={10.4218/etrij.2020-0156},
abstract={5G AgiLe and fLexible integration of SaTellite And cellulaR (5G-ALLSTAR) is a Korea-Europe (KR-EU) collaborative project for developing multi-connectivity (MC) technologies that integrate cellular and satellite networks to provide seamless, reliable, and ubiquitous broadband communication services and improve service continuity for 5G and beyond. The main scope of this project entails the prototype development of a millimeter-wave 5G New Radio (NR)-based cellular system, an investigation of the feasibility of an NR-based satellite system and its integration with cellular systems, and a study of spectrum sharing and interference management techniques for MC. This article reviews recent research activities and presents preliminary results and a plan for the proof of concept (PoC) of three representative use cases (UCs) and one joint KR-EU UC. The feasibility of each UC and superiority of the developed technologies will be validated with key performance indicators using corresponding PoC platforms. The final achievements of the project are expected to eventually contribute to the technical evolution of 5G, which will pave the road for next-generation communications. © 2020 ETRI},
author_keywords={5G-ALLSTAR;  millimeter-wave;  multi-connectivity;  new radio;  satellite communications;  vehicular communications},
keywords={Benchmarking;  Millimeter waves;  Mobile telecommunication systems;  Satellites, Broadband Communication;  Collaborative projects;  Flexible integration;  Interference management;  Key performance indicators;  Prototype development;  Service continuity;  Technical evolution, 5G mobile communication systems},
document_type={Article},
}

5G AgiLe and fLexible integration of SaTellite And cellulaR (5G-ALLSTAR) is a Korea-Europe (KR-EU) collaborative project for developing multi-connectivity (MC) technologies that integrate cellular and satellite networks to provide seamless, reliable, and ubiquitous broadband communication services and improve service continuity for 5G and beyond. The main scope of this project entails the prototype development of a millimeter-wave 5G New Radio (NR)-based cellular system, an investigation of the feasibility of an NR-based satellite system and its integration with cellular systems, and a study of spectrum sharing and interference management techniques for MC. This article reviews recent research activities and presents preliminary results and a plan for the proof of concept (PoC) of three representative use cases (UCs) and one joint KR-EU UC. The feasibility of each UC and superiority of the developed technologies will be validated with key performance indicators using corresponding PoC platforms. The final achievements of the project are expected to eventually contribute to the technical evolution of 5G, which will pave the road for next-generation communications. © 2020 ETRI

@ARTICLE{CalvaneseStrinati2020643,
author={Calvanese Strinati, E. and Barbarossa, S. and Choi, T. and Pietrabissa, A. and Giuseppi, A. and De Santis, E. and Vidal, J. and Becvar, Z. and Haustein, T. and Cassiau, N. and Costanzo, F. and Kim, J. and Kim, I.},
title={6G in the sky: On-demand intelligence at the edge of 3D networks (Invited paper)},
journal={ETRI Journal},
year={2020},
volume={42},
number={5},
pages={643-657},
doi={10.4218/etrij.2020-0205},
abstract={Sixth generation will exploit satellite, aerial, and terrestrial platforms jointly to improve radio access capability and unlock the support of on-demand edge cloud services in three-dimensional (3D) space, by incorporating mobile edge computing (MEC) functionalities on aerial platforms and low-orbit satellites. This will extend the MEC support to devices and network elements in the sky and forge a space-borne MEC, enabling intelligent, personalized, and distributed on-demand services. End users will experience the impression of being surrounded by a distributed computer, fulfilling their requests with apparently zero latency. In this paper, we consider an architecture that provides communication, computation, and caching (C3) services on demand, anytime, and everywhere in 3D space, integrating conventional ground (terrestrial) base stations and flying (non-terrestrial) nodes. Given the complexity of the overall network, the C3 resources and management of aerial devices need to be jointly orchestrated via artificial intelligence-based algorithms, exploiting virtualized network functions dynamically deployed in a distributed manner across terrestrial and non-terrestrial nodes. © 2020 ETRI},
author_keywords={3D connectivity;  3D services;  5G;  6G;  B5G;  high-altitude platform stations;  mobile edge computing;  non-terrestrial communications;  satellite;  unmanned aerial vehicle},
keywords={Antennas;  Artificial intelligence;  Orbits;  Space platforms;  User experience, Aerial platform;  Low-orbit satellites;  Network element;  Network functions;  On-demand services;  Overall networks;  Services on demand;  Three-dimensional (3D) space, Network function virtualization},
document_type={Article},
}

Sixth generation will exploit satellite, aerial, and terrestrial platforms jointly to improve radio access capability and unlock the support of on-demand edge cloud services in three-dimensional (3D) space, by incorporating mobile edge computing (MEC) functionalities on aerial platforms and low-orbit satellites. This will extend the MEC support to devices and network elements in the sky and forge a space-borne MEC, enabling intelligent, personalized, and distributed on-demand services. End users will experience the impression of being surrounded by a distributed computer, fulfilling their requests with apparently zero latency. In this paper, we consider an architecture that provides communication, computation, and caching (C3) services on demand, anytime, and everywhere in 3D space, integrating conventional ground (terrestrial) base stations and flying (non-terrestrial) nodes. Given the complexity of the overall network, the C3 resources and management of aerial devices need to be jointly orchestrated via artificial intelligence-based algorithms, exploiting virtualized network functions dynamically deployed in a distributed manner across terrestrial and non-terrestrial nodes. © 2020 ETRI

@CONFERENCE{Ornatelli2020292,
author={Ornatelli, A. and Giuseppi, A. and Suraci, V. and Tortorelli, A.},
title={User-aware centralized resource allocation in heterogeneous networks},
journal={2020 28th Mediterranean Conference on Control and Automation, MED 2020},
year={2020},
pages={292-298},
doi={10.1109/MED48518.2020.9183080},
art_number={9183080},
abstract={In the last two years, in Europe, 5G networks and services proliferated. The integration of 5G networks with other radio access networks is considered one of the key enablers for matching the challenging 5G Quality of Service requirements. In particular, the integration with high throughput satellites promises to increase the network performances in terms of resilience and Quality of Service. The present work addresses this problem and presents a user-aware resource allocation methodology for heterogeneous networks. Said methodology is articulated in two-steps: at first, the Analytical Hierarchy Process is used for deciding the network over which traffic is steered and, then, a Cooperative Game for allocating resources within the network is set up. Simulations are presented for validating the proposed approach. © 2020 IEEE.},
author_keywords={Analytical hierarchy process;  Cooperative differential games;  Multi-connectivity;  Network selection;  Traffic steering},
document_type={Conference Paper},
}

In the last two years, in Europe, 5G networks and services proliferated. The integration of 5G networks with other radio access networks is considered one of the key enablers for matching the challenging 5G Quality of Service requirements. In particular, the integration with high throughput satellites promises to increase the network performances in terms of resilience and Quality of Service. The present work addresses this problem and presents a user-aware resource allocation methodology for heterogeneous networks. Said methodology is articulated in two-steps: at first, the Analytical Hierarchy Process is used for deciding the network over which traffic is steered and, then, a Cooperative Game for allocating resources within the network is set up. Simulations are presented for validating the proposed approach. © 2020 IEEE.

@CONFERENCE{Giuseppi2020746,
author={Giuseppi, A. and Pietrabissa, A. and Liberati, F. and Di Giorgio, A.},
title={Controlled optimal black start procedures in smart grids for service restoration in presence of electrical storage systems},
journal={2020 28th Mediterranean Conference on Control and Automation, MED 2020},
year={2020},
pages={746-751},
doi={10.1109/MED48518.2020.9183176},
art_number={9183176},
abstract={This paper presents an optimisation problem to determine the optimal reclosure order of remotely operable switches deployed in a smart grid consisting in a distribution network equipped with one or more Energy Storage Systems (ESS). The proposed solution integrates nonlinear real and reactive power flow equations, by reconducting them to a set of conic constraints, together with several network operator requirements, such as network radiality and ampacity limits. A numerical simulation validates the approach and concludes the work. © 2020 IEEE.},
author_keywords={Controlled Black Start;  Mixed Integer Programming;  Power Network Resiliency;  Service Restoration},
document_type={Conference Paper},
}

This paper presents an optimisation problem to determine the optimal reclosure order of remotely operable switches deployed in a smart grid consisting in a distribution network equipped with one or more Energy Storage Systems (ESS). The proposed solution integrates nonlinear real and reactive power flow equations, by reconducting them to a set of conic constraints, together with several network operator requirements, such as network radiality and ampacity limits. A numerical simulation validates the approach and concludes the work. © 2020 IEEE.

@ARTICLE{Liberati20203586,
author={Liberati, F. and Di Giorgio, A. and Giuseppi, A. and Pietrabissa, A. and Priscoli, F.D.},
title={Efficient and Risk-Aware Control of Electricity Distribution Grids},
journal={IEEE Systems Journal},
year={2020},
volume={14},
number={3},
pages={3586-3597},
doi={10.1109/JSYST.2020.2965633},
art_number={8966461},
abstract={This article presents an economic model predictive control (EMPC) algorithm for reducing losses and increasing the resilience of medium-voltage electricity distribution grids characterized by high penetration of renewable energy sources and possibly subject to natural or malicious adverse events. The proposed control system optimizes grid operations through network reconfiguration, control of distributed energy storage systems (ESSs), and on-load tap changers. The core of the EMPC algorithm is a nonconvex optimization problem integrating the ESSs dynamics, the topological and power technical constraints of the grid, and the modeling of the cascading effects of potential adverse events. An equivalent (i.e., having the same optimal solution) proxy of the nonconvex problem is proposed to make the solution more tractable. Simulations performed on a 16-bus test distribution network validate the proposed control strategy. © 2007-2012 IEEE.},
author_keywords={Energy storage systems (ESSs);  model predictive control (MPC);  network reconfiguration;  resilient control;  smart grids},
keywords={Data storage equipment;  Electric energy storage;  Electric utilities;  Model predictive control;  Renewable energy resources;  Voltage control, Control strategies;  Distributed energy storage systems;  Electricity distribution;  Network re-configuration;  Nonconvex optimization problem;  On- load tap changers;  Renewable energy source;  Technical constraints, Electric power transmission networks},
document_type={Article},
}

This article presents an economic model predictive control (EMPC) algorithm for reducing losses and increasing the resilience of medium-voltage electricity distribution grids characterized by high penetration of renewable energy sources and possibly subject to natural or malicious adverse events. The proposed control system optimizes grid operations through network reconfiguration, control of distributed energy storage systems (ESSs), and on-load tap changers. The core of the EMPC algorithm is a nonconvex optimization problem integrating the ESSs dynamics, the topological and power technical constraints of the grid, and the modeling of the cascading effects of potential adverse events. An equivalent (i.e., having the same optimal solution) proxy of the nonconvex problem is proposed to make the solution more tractable. Simulations performed on a 16-bus test distribution network validate the proposed control strategy. © 2007-2012 IEEE.

@ARTICLE{Giuseppi2020755,
author={Giuseppi, A. and Pietrabissa, A.},
title={Chance-Constrained Control with Lexicographic Deep Reinforcement Learning},
journal={IEEE Control Systems Letters},
year={2020},
volume={4},
number={3},
pages={755-760},
doi={10.1109/LCSYS.2020.2979635},
art_number={9031720},
abstract={This letter proposes a lexicographic Deep Reinforcement Learning (DeepRL)-based approach to chance-constrained Markov Decision Processes, in which the controller seeks to ensure that the probability of satisfying the constraint is above a given threshold. Standard DeepRL approaches require i) the constraints to be included as additional weighted terms in the cost function, in a multi-objective fashion, and ii) the tuning of the introduced weights during the training phase of the Deep Neural Network (DNN) according to the probability thresholds. The proposed approach, instead, requires to separately train one constraint-free DNN and one DNN associated to each constraint and then, at each time-step, to select which DNN to use depending on the system observed state. The presented solution does not require any hyper-parameter tuning besides the standard DNN ones, even if the probability thresholds changes. A lexicographic version of the well-known DeepRL algorithm DQN is also proposed and validated via simulations. © 2017 IEEE.},
author_keywords={constrained control;  deep reinforcement learning;  Markov decision processes},
keywords={Cost functions;  Deep neural networks;  Markov processes;  Reinforcement learning, Chance-constrained;  Chance-constrained controls;  Constrained controls;  Hyper-parameter;  Markov Decision Processes;  Multi objective;  Probability threshold;  Training phase, Deep learning},
document_type={Article},
}

This letter proposes a lexicographic Deep Reinforcement Learning (DeepRL)-based approach to chance-constrained Markov Decision Processes, in which the controller seeks to ensure that the probability of satisfying the constraint is above a given threshold. Standard DeepRL approaches require i) the constraints to be included as additional weighted terms in the cost function, in a multi-objective fashion, and ii) the tuning of the introduced weights during the training phase of the Deep Neural Network (DNN) according to the probability thresholds. The proposed approach, instead, requires to separately train one constraint-free DNN and one DNN associated to each constraint and then, at each time-step, to select which DNN to use depending on the system observed state. The presented solution does not require any hyper-parameter tuning besides the standard DNN ones, even if the probability thresholds changes. A lexicographic version of the well-known DeepRL algorithm DQN is also proposed and validated via simulations. © 2017 IEEE.

@ARTICLE{Valensise2020,
author={Valensise, C.M. and Giuseppi, A. and Vernuccio, F. and De La Cadena, A. and Cerullo, G. and Polli, D.},
title={Removing non-resonant background from CARS spectra via deep learning},
journal={APL Photonics},
year={2020},
volume={5},
number={6},
doi={10.1063/5.0007821},
art_number={061305},
abstract={Broadband Coherent Anti-Stokes Raman Scattering (B-CARS) is a powerful label-free nonlinear spectroscopy technique allowing one to measure the full vibrational spectrum of molecules and solids. B-CARS spectra, however, suffer from the presence of a spurious signal, called non-resonant background (NRB), which interferes with the resonant vibrational one, distorting the line shapes and degrading the chemical information. While several numerical techniques are available to remove this unwanted contribution and extract the resonant vibrational signal of interest, they all require the user's intervention and sensitively depend on the spectral shape of the NRB, which needs to be measured independently. In this work, we present a novel approach to remove NRB from B-CARS spectra based on deep learning. Thanks to the high generalization capability offered by the deep architecture of the designed neural network, trained through realistic simulated spectra, our fully automated model is able to process B-CARS spectra in real time and independently of the detailed shape of the NRB spectrum. This results in fast extraction of vibrational spectra without requiring user intervention or the measurement of reference spectra. We expect that this model will significantly simplify and speed-up the analysis of B-CARS spectra in spectroscopy and microscopy. © 2020 Author(s).},
keywords={Coherent scattering;  Raman scattering;  Raman spectroscopy;  Vibrational spectra, Chemical information;  Coherent anti Stokes Raman scattering;  Deep architectures;  Generalization capability;  Nonlinear spectroscopy;  Numerical techniques;  Reference spectrum;  Vibrational signals, Deep learning},
document_type={Article},
}

Broadband Coherent Anti-Stokes Raman Scattering (B-CARS) is a powerful label-free nonlinear spectroscopy technique allowing one to measure the full vibrational spectrum of molecules and solids. B-CARS spectra, however, suffer from the presence of a spurious signal, called non-resonant background (NRB), which interferes with the resonant vibrational one, distorting the line shapes and degrading the chemical information. While several numerical techniques are available to remove this unwanted contribution and extract the resonant vibrational signal of interest, they all require the user's intervention and sensitively depend on the spectral shape of the NRB, which needs to be measured independently. In this work, we present a novel approach to remove NRB from B-CARS spectra based on deep learning. Thanks to the high generalization capability offered by the deep architecture of the designed neural network, trained through realistic simulated spectra, our fully automated model is able to process B-CARS spectra in real time and independently of the detailed shape of the NRB spectrum. This results in fast extraction of vibrational spectra without requiring user intervention or the measurement of reference spectra. We expect that this model will significantly simplify and speed-up the analysis of B-CARS spectra in spectroscopy and microscopy. © 2020 Author(s).

@CONFERENCE{Germana20201330,
author={Germana, R. and Giuseppi, A. and Di Giorgio, A.},
title={Ensuring the Stability of Power Systems Against Dynamic Load Altering Attacks: A Robust Control Scheme Using Energy Storage Systems},
journal={European Control Conference 2020, ECC 2020},
year={2020},
pages={1330-1335},
art_number={9143620},
abstract={This paper presents a robust protection scheme to protect the power transmission network against a class of feedback-based attacks referred in the literature as "Dynamic Load Altering Attacks" (D-LAAs). The proposed scheme envisages the usage of Energy Storage Systems (ESSs) to avoid the destabilising effects that a malicious state feedback has on the power network generators. The methodologies utilised are based on results from polytopic uncertain systems, invariance theory and Lyapunov arguments. Numerical simulations on a test scenario validate the proposed approach. © 2020 EUCA.},
author_keywords={Dynamic Load Altering Attacks;  Electric Energy Storage Systems;  Robust Control},
keywords={Data storage equipment;  Dynamic loads;  Electric equipment protection;  Electric power system protection;  Electric power transmission networks;  Energy storage;  Robust control;  State feedback, Energy storage systems;  Energy Storage Systems (ESSs);  Feed-back based;  Invariance theory;  Polytopic uncertain systems;  Protection schemes;  Robust control scheme;  Stability of power system, Electric power system stability},
document_type={Conference Paper},
}

This paper presents a robust protection scheme to protect the power transmission network against a class of feedback-based attacks referred in the literature as "Dynamic Load Altering Attacks" (D-LAAs). The proposed scheme envisages the usage of Energy Storage Systems (ESSs) to avoid the destabilising effects that a malicious state feedback has on the power network generators. The methodologies utilised are based on results from polytopic uncertain systems, invariance theory and Lyapunov arguments. Numerical simulations on a test scenario validate the proposed approach. © 2020 EUCA.

@CONFERENCE{Priscoli2020595,
author={Priscoli, F.D. and Giuseppi, A. and Liberati, F. and Pietrabissa, A.},
title={Traffic Steering and Network Selection in 5G Networks based on Reinforcement Learning},
journal={European Control Conference 2020, ECC 2020},
year={2020},
pages={595-601},
art_number={9143837},
abstract={This paper presents a controller for the problem of Network Selection in 5G Networks, based on Reinforcement Learning. The problem of Network Selection and Traffic Steering is modeled as a Markov Decision Process and a Q- Learning based control solution is designed to meet 5G requirements, such as Quality of Experience (QoE) maximization, Quality of Service (QoS) assurance and load balancing. Numerical simulations preliminarily validate the proposed approach on a simulated scenario considered in the European project H2020 5G-ALLSTAR. © 2020 EUCA.},
keywords={Markov processes;  Quality control;  Quality of service;  Queueing networks;  Reinforcement learning, Control solutions;  European project;  G-networks;  Markov Decision Processes;  Network selection;  Q-learning;  Quality of experience (QoE), 5G mobile communication systems},
document_type={Conference Paper},
}

This paper presents a controller for the problem of Network Selection in 5G Networks, based on Reinforcement Learning. The problem of Network Selection and Traffic Steering is modeled as a Markov Decision Process and a Q- Learning based control solution is designed to meet 5G requirements, such as Quality of Experience (QoE) maximization, Quality of Service (QoS) assurance and load balancing. Numerical simulations preliminarily validate the proposed approach on a simulated scenario considered in the European project H2020 5G-ALLSTAR. © 2020 EUCA.

@CONFERENCE{Giuseppi2020,
author={Giuseppi, A. and De Santis, E. and Delli Priscoli, F. and Won, S.H. and Choi, T. and Pietrabissa, A.},
title={Network Selection in 5G Networks Based on Markov Games and Friend-or-Foe Reinforcement Learning},
journal={2020 IEEE Wireless Communications and Networking Conference Workshops, WCNCW 2020 - Proceedings},
year={2020},
doi={10.1109/WCNCW48565.2020.9124723},
art_number={9124723},
abstract={This paper presents a control solution for the optimal network selection problem in 5G heterogeneous networks. The control logic proposed is based on multi-agent Friend-or-Foe Q-Learning, allowing the design of a distributed control architecture that sees the various access points compete for the allocation of the connection requests. Numerical simulations validate conceptually the approach, developed in the scope of the EU-Korea project 5G-ALLSTAR. © 2020 IEEE.},
author_keywords={5G;  Markov Games;  Multi-Agent Reinforcement Learning;  Network Selection},
keywords={Heterogeneous networks;  Multi agent systems;  Queueing networks;  Reinforcement learning, Access points;  Control logic;  Control solutions;  Distributed control architectures;  Friend or Foe-Q learning;  Markov games;  Network selection;  Optimal networks, 5G mobile communication systems},
document_type={Conference Paper},
}

This paper presents a control solution for the optimal network selection problem in 5G heterogeneous networks. The control logic proposed is based on multi-agent Friend-or-Foe Q-Learning, allowing the design of a distributed control architecture that sees the various access points compete for the allocation of the connection requests. Numerical simulations validate conceptually the approach, developed in the scope of the EU-Korea project 5G-ALLSTAR. © 2020 IEEE.

@CONFERENCE{Kim2020,
author={Kim, J. and Casati, G. and Pietrabissa, A. and Giuseppi, A. and Calvanese Strinati, E. and Cassiau, N. and Noh, G. and Chung, H. and Kim, I. and Thary, M. and Houssin, J.-M. and Pigni, F. and Colombero, S. and Dal Zotto, P. and Raschkowski, L. and Jaeckel, S.},
title={5G-ALLSTAR: An Integrated Satellite-Cellular System for 5G and beyond},
journal={2020 IEEE Wireless Communications and Networking Conference Workshops, WCNCW 2020 - Proceedings},
year={2020},
doi={10.1109/WCNCW48565.2020.9124751},
art_number={9124751},
abstract={This paper provides an overview of recent research activities of the 5G AgiLe and fLexible integration of SaTellite And cellulaR (5G-ALLSTAR) project which aims to develop Multi-Connectivity technology that integrates the cellular and satellite networks to provide seamless, reliable and ubiquitous broadband services. 5G-ALLSTAR also entails developing millimeter-wave (mmWave) 5G New Radio (NR)-based cellular access system and investigating the feasibility of NR-based satellite access for providing broadband and reliable 5G services. In addition, spectrum sharing between cellular and satellite networks is studied. With all the technologies developed, 5G-ALLSTAR will showcase the first fully integrated satellite and cellular prototype system for 5G and beyond 5G (B5G) services at a big event (e.g., sporting event like Roland-Garros) in 2021. This paper also provides a preliminary techno-economic analysis on potential use cases targeting vertical markets, and introduces recent standardization activities of relevance. © 2020 IEEE.},
author_keywords={5G-ALLSTAR;  millimeter wave;  multi-connectivity;  New Radio;  satellite communications;  vehicular communications},
keywords={Economic analysis;  Millimeter waves;  Satellites, Broadband service;  Flexible integration;  Millimeter waves (mmwave);  Prototype system;  Recent researches;  Satellite access;  Satellite network;  Techno-Economic analysis, 5G mobile communication systems},
document_type={Conference Paper},
}

This paper provides an overview of recent research activities of the 5G AgiLe and fLexible integration of SaTellite And cellulaR (5G-ALLSTAR) project which aims to develop Multi-Connectivity technology that integrates the cellular and satellite networks to provide seamless, reliable and ubiquitous broadband services. 5G-ALLSTAR also entails developing millimeter-wave (mmWave) 5G New Radio (NR)-based cellular access system and investigating the feasibility of NR-based satellite access for providing broadband and reliable 5G services. In addition, spectrum sharing between cellular and satellite networks is studied. With all the technologies developed, 5G-ALLSTAR will showcase the first fully integrated satellite and cellular prototype system for 5G and beyond 5G (B5G) services at a big event (e.g., sporting event like Roland-Garros) in 2021. This paper also provides a preliminary techno-economic analysis on potential use cases targeting vertical markets, and introduces recent standardization activities of relevance. © 2020 IEEE.

@ARTICLE{Giuseppi2020,
author={Giuseppi, A. and Pietrabissa, A.},
title={Wardrop equilibrium in discrete-time selfish routing with time-varying bounded delays},
journal={IEEE Transactions on Automatic Control},
year={2020},
doi={10.1109/TAC.2020.2981906},
art_number={2981906},
abstract={This paper presents a multi-commodity, discrete-time, distributed and non-cooperative routing algorithm, which is proved to converge to an equilibrium in the presence of heterogeneous, unknown, time-varying but bounded delays. Under mild assumptions on the latency functions which describe the cost associated to the network paths, two algorithms are proposed: the former assumes that each commodity relies only on measurements of the latencies associated to its own paths; the latter assumes that each commodity has (at least indirectly) access to the measures of the latencies of all the network paths. Both algorithms are proven to drive the system state to an invariant set which approximates and contains the Wardrop equilibrium, defined as a network state in which no traffic flow over the network paths can improve its routing unilaterally, with the latter achieving a better reconstruction of the Wardrop equilibrium. Numerical simulations show the effectiveness of the proposed approach. © 2020 Institute of Electrical and Electronics Engineers Inc.. All rights reserved.},
author_keywords={LaSalle's invariance principle;  Selfish routing;  Time-delay systems;  Wardrop equilibrium},
keywords={Mathematical models, Bounded delays;  Latency function;  Measurements of;  Multi-commodity;  Network state;  Non-cooperative;  Selfish routing;  Wardrop equilibrium, Control systems},
document_type={Article},
}

This paper presents a multi-commodity, discrete-time, distributed and non-cooperative routing algorithm, which is proved to converge to an equilibrium in the presence of heterogeneous, unknown, time-varying but bounded delays. Under mild assumptions on the latency functions which describe the cost associated to the network paths, two algorithms are proposed: the former assumes that each commodity relies only on measurements of the latencies associated to its own paths; the latter assumes that each commodity has (at least indirectly) access to the measures of the latencies of all the network paths. Both algorithms are proven to drive the system state to an invariant set which approximates and contains the Wardrop equilibrium, defined as a network state in which no traffic flow over the network paths can improve its routing unilaterally, with the latter achieving a better reconstruction of the Wardrop equilibrium. Numerical simulations show the effectiveness of the proposed approach. © 2020 Institute of Electrical and Electronics Engineers Inc.. All rights reserved.

@CONFERENCE{Choi202040,
author={Choi, T. and Won, S.H. and Giuseppi, A. and Pietrabissa, A. and Kwon, S.},
title={Management and Orchestration Architecture for Integrated Access of Satellite and Terrestrial in 5G},
journal={International Conference on Information Networking},
year={2020},
volume={2020-January},
pages={40-45},
doi={10.1109/ICOIN48656.2020.9016484},
art_number={9016484},
abstract={Multi-RAT access network, or heterogeneous access network, is considered to be the key enabling technology to satisfy the 5G requirements, such as high data rate, ultra-low latency and reliability. To make efficient use of all the available network resources, various research activities on multi-connectivity have been proposed to simultaneously connect, steer, and orchestrate across multiple different radio access technologies. Standardization of the management and orchestration of multi-connectivity environment, however, has just been initiated, thus further research and development is required. This paper proposes a novel management and orchestration architecture for integrated access of satellite and terrestrial in 5G. It especially focuses on the traffic steering and load-balancing of heterogeneous multi-RAT access environment. © 2020 IEEE.},
author_keywords={load-balancing;  management and orchestration;  multi-connectivity;  QoS/QoE management;  traffic steering},
keywords={5G mobile communication systems;  Network architecture;  Rats;  Resource allocation, Enabling technologies;  Heterogeneous access;  Integrated access;  multi-connectivity;  QoS/QoE;  Radio access technologies;  Research activities;  Research and development, Research and development management},
document_type={Conference Paper},
}

Multi-RAT access network, or heterogeneous access network, is considered to be the key enabling technology to satisfy the 5G requirements, such as high data rate, ultra-low latency and reliability. To make efficient use of all the available network resources, various research activities on multi-connectivity have been proposed to simultaneously connect, steer, and orchestrate across multiple different radio access technologies. Standardization of the management and orchestration of multi-connectivity environment, however, has just been initiated, thus further research and development is required. This paper proposes a novel management and orchestration architecture for integrated access of satellite and terrestrial in 5G. It especially focuses on the traffic steering and load-balancing of heterogeneous multi-RAT access environment. © 2020 IEEE.

@ARTICLE{Tortorelli202074,
author={Tortorelli, A. and Fiaschetti, A. and Giuseppi, A. and Suraci, V. and Germanà, R. and Priscoli, F.D.},
title={A security metric for assessing the security level of critical infrastructures},
journal={International Journal of Critical Computer-Based Systems},
year={2020},
volume={10},
number={1},
pages={74-94},
doi={10.1504/IJCCBS.2020.108685},
abstract={The deep integration between the cyber and physical domains in complex systems make very challenging the security evaluation process, as security itself is more of a concept (i.e., a subjective property) than a quantifiable characteristic. Traditional security assessing mostly relies on the personal skills of security experts, often based on best practices and personal experience. The present work is aimed at defining a security metric allowing evaluators to assess the security level of complex cyber-physical systems (CPSs), as critical infrastructures, in a holistic, consistent and repeatable way. To achieve this result, the mathematical framework provided by the open source security testing methodology manual (OSSTMM) is used as the backbone of the new security metric, since it allows to provide security indicators capturing, in a non-biased way, the security level of a system. Several concepts, as component lifecycle, vulnerability criticality and damage potential – effort ratio are embedded in the new security metric framework, developed in the scope of the H2020 project ATENA. Copyright © 2020 Inderscience Enterprises Ltd.},
author_keywords={CPSs;  Critical infrastructures;  Cyber-physical security;  Cyber-physical systems;  Security metrics},
document_type={Article},
}

The deep integration between the cyber and physical domains in complex systems make very challenging the security evaluation process, as security itself is more of a concept (i.e., a subjective property) than a quantifiable characteristic. Traditional security assessing mostly relies on the personal skills of security experts, often based on best practices and personal experience. The present work is aimed at defining a security metric allowing evaluators to assess the security level of complex cyber-physical systems (CPSs), as critical infrastructures, in a holistic, consistent and repeatable way. To achieve this result, the mathematical framework provided by the open source security testing methodology manual (OSSTMM) is used as the backbone of the new security metric, since it allows to provide security indicators capturing, in a non-biased way, the security level of a system. Several concepts, as component lifecycle, vulnerability criticality and damage potential – effort ratio are embedded in the new security metric framework, developed in the scope of the H2020 project ATENA. Copyright © 2020 Inderscience Enterprises Ltd.

@ARTICLE{Liberati20197015,
author={Liberati, F. and Giorgio, A.D. and Giuseppi, A. and Pietrabissa, A. and Habib, E. and Martirano, L.},
title={Joint Model Predictive Control of Electric and Heating Resources in a Smart Building},
journal={IEEE Transactions on Industry Applications},
year={2019},
volume={55},
number={6},
pages={7015-7027},
doi={10.1109/TIA.2019.2932954},
art_number={8786121},
abstract={The new challenge in power systems design and operation is to organize and control smart micro grids supplying aggregation of users and special loads as electric vehicles charging stations. The presence of renewable and storage can help the optimal operation only if a good control manages all the elements of the grid. New models of green buildings and energy communities are proposed. For a real application they need an appropriate and advanced power system equipped with a building automation control system. This article presents an economic model predictive control approach to the problem of managing the electric and heating resources in a smart building in a coordinated way, for the purpose of achieving in real time nearly zero energy consumption and automated participation to demand response programs. The proposed control, leveraging a mixed integer quadratic programming problem, allows to meet manifold thermal and electric users' requirements and react to inbound demand response signals, while still guaranteeing stable operation of the building's electric and thermal storage equipment. The simulation results, performed for a real case study in Italy, highlight the peculiarities of the proposed approach in the joint handling of electric and thermal building flexibility. © 1972-2012 IEEE.},
author_keywords={Demand side management;  economic model predictive control (EMPC);  heating systems;  smart buildings},
keywords={Demand side management;  Electric control equipment;  Electric utilities;  Energy utilization;  Heat storage;  Integer programming;  Intelligent buildings;  Model predictive control;  Quadratic programming;  Vehicle-to-grid;  Zero energy buildings, Building automation;  Demand response programs;  Economic modeling;  Heating system;  Mixed integer quadratic programming;  Real applications;  Thermal buildings;  Thermal storage equipments, Electric power system control},
document_type={Article},
}

The new challenge in power systems design and operation is to organize and control smart micro grids supplying aggregation of users and special loads as electric vehicles charging stations. The presence of renewable and storage can help the optimal operation only if a good control manages all the elements of the grid. New models of green buildings and energy communities are proposed. For a real application they need an appropriate and advanced power system equipped with a building automation control system. This article presents an economic model predictive control approach to the problem of managing the electric and heating resources in a smart building in a coordinated way, for the purpose of achieving in real time nearly zero energy consumption and automated participation to demand response programs. The proposed control, leveraging a mixed integer quadratic programming problem, allows to meet manifold thermal and electric users' requirements and react to inbound demand response signals, while still guaranteeing stable operation of the building's electric and thermal storage equipment. The simulation results, performed for a real case study in Italy, highlight the peculiarities of the proposed approach in the joint handling of electric and thermal building flexibility. © 1972-2012 IEEE.

@CONFERENCE{Giuseppi20193365,
author={Giuseppi, A. and De Santis, E. and Di Giorgio, A.},
title={Model predictive control of energy storage systems for power regulation in electricity distribution networks},
journal={Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics},
year={2019},
volume={2019-October},
pages={3365-3370},
doi={10.1109/SMC.2019.8914059},
art_number={8914059},
abstract={This paper proposes a control strategy for an Energy Storage System (ESS) installed on a secondary substation of an electricity distribution line in order to mitigate power variations with respect to the day-ahead planning caused by renewable energy sources on the distribution line.In particular, the aim of the controller is to keep the power profile of at primary substations close to a reference profile foreseen on a day-ahead basis while guaranteeing the stable operation of its ESS, in terms of their state-of-charge dynamics. The inclusion of the ESS contribution to the network operation is attained by the integration of properly defined power flow constraints in a Model Predictive Control Framework. The proposed approach has been validated through numerical simulations, representative of real operative scenarios. © 2019 IEEE.},
keywords={Data storage equipment;  Electric energy storage;  Electric load flow;  Electric power system control;  Electric utilities;  Energy policy;  Model predictive control;  Renewable energy resources, Control strategies;  Electricity distribution;  Electricity distribution networks;  Energy storage systems;  Network operations;  Power regulation;  Renewable energy source;  Stable operation, Electric power system planning},
document_type={Conference Paper},
}

This paper proposes a control strategy for an Energy Storage System (ESS) installed on a secondary substation of an electricity distribution line in order to mitigate power variations with respect to the day-ahead planning caused by renewable energy sources on the distribution line.In particular, the aim of the controller is to keep the power profile of at primary substations close to a reference profile foreseen on a day-ahead basis while guaranteeing the stable operation of its ESS, in terms of their state-of-charge dynamics. The inclusion of the ESS contribution to the network operation is attained by the integration of properly defined power flow constraints in a Model Predictive Control Framework. The proposed approach has been validated through numerical simulations, representative of real operative scenarios. © 2019 IEEE.

@CONFERENCE{Giorgio20193216,
author={Giorgio, A.D. and Giuseppi, A. and Germana, R. and Liberati, F.},
title={Decentralised model predictive control of electric vehicles charging},
journal={Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics},
year={2019},
volume={2019-October},
pages={3216-3222},
doi={10.1109/SMC.2019.8914040},
art_number={8914040},
abstract={This paper presents a decentralised control strategy for the management of simultaneous charging sessions of electric vehicles. The proposed approach is based on the model predictive control methodology and the Lagrangian decomposition of the constrained optimization problem which is solved at each sampling time. This strategy allows the computation of the charging profiles in a decentralised way, with limited information exchange between the electric vehicles. The simulation results show the potential of the proposed approach in relation to the problem of shaving the aggregated power withdrawal from the electricity distribution grid, while still satisfying drivers' preferences for charging. © 2019 IEEE.},
keywords={Charging (batteries);  Constrained optimization;  Decentralized control;  Electric utilities;  Electric vehicles, Charging profiles;  Constrained optimi-zation problems;  Decentralised;  Decentralised control;  Electricity distribution;  Lagrangian decomposition;  Limited information;  Sampling time, Model predictive control},
document_type={Conference Paper},
}

This paper presents a decentralised control strategy for the management of simultaneous charging sessions of electric vehicles. The proposed approach is based on the model predictive control methodology and the Lagrangian decomposition of the constrained optimization problem which is solved at each sampling time. This strategy allows the computation of the charging profiles in a decentralised way, with limited information exchange between the electric vehicles. The simulation results show the potential of the proposed approach in relation to the problem of shaving the aggregated power withdrawal from the electricity distribution grid, while still satisfying drivers' preferences for charging. © 2019 IEEE.

@ARTICLE{Giuseppi2019221,
author={Giuseppi, A. and Pietrabissa, A. and Cilione, S. and Galvagni, L.},
title={Feedback linearization-based satellite attitude control with a life-support device without communications},
journal={Control Engineering Practice},
year={2019},
volume={90},
pages={221-230},
doi={10.1016/j.conengprac.2019.06.020},
abstract={This paper develops a control strategy for a life-support device to be attached to an orbiting satellite to extend its operational life. The objective is met in such a way that the original satellite keeps operating without communications between the two systems (also valuable for energy efficiency). The case in which the original satellite is equipped with a feedback-linearization based controller is considered and the control law for the life-support is developed with the same methodology, obtaining a compensating control which recovers the performance of the original control strategy. Simulations validate the approach considering a real case study in various scenarios. © 2019 Elsevier Ltd},
author_keywords={Feedback linearization;  Model predictive control;  Satellite systems;  Spacecraft attitude control},
keywords={Artificial life;  Attitude control;  Control theory;  Energy efficiency;  Model predictive control;  Orbits;  Satellites;  Space flight, Compensating control;  Control strategies;  Life supports;  Operational life;  Orbiting satellites;  Satellite attitude control;  Satellite system;  Spacecraft attitude control, Feedback linearization},
document_type={Article},
}

This paper develops a control strategy for a life-support device to be attached to an orbiting satellite to extend its operational life. The objective is met in such a way that the original satellite keeps operating without communications between the two systems (also valuable for energy efficiency). The case in which the original satellite is equipped with a feedback-linearization based controller is considered and the control law for the life-support is developed with the same methodology, obtaining a compensating control which recovers the performance of the original control strategy. Simulations validate the approach considering a real case study in various scenarios. © 2019 Elsevier Ltd

@CONFERENCE{Giuseppi201950,
author={Giuseppi, A. and Tortorelli, A. and Germana, R. and Liberati, F. and Fiaschetti, A.},
title={Securing cyber-physical systems: An optimization framework based on OSSTMM and genetic algorithms},
journal={27th Mediterranean Conference on Control and Automation, MED 2019 - Proceedings},
year={2019},
pages={50-56},
doi={10.1109/MED.2019.8798506},
art_number={8798506},
abstract={This paper presents an optimization framework, based on Genetic Algorithms, for the control of the 'security level' of a Cyber-Physical System (CPS). The security level is a quantity that has been studied in several industrial standards, among which we selected the Open Source Security Testing Methodology Manual (OSSTMM). The proposed optimization solution is validated on scenarios representative of real operations of a security evaluator, and numerical simulations report the performances obtained by the algorithm. © 2019 IEEE.},
author_keywords={Cyber-Phisycal systems;  Optimal Planning;  Security},
keywords={Cyber Physical System;  Genetic algorithms, Cyber-physical systems (CPS);  Industrial standards;  Optimal planning;  Optimization framework;  Optimization solution;  Security;  Security level;  Security testing, Embedded systems},
document_type={Conference Paper},
}

This paper presents an optimization framework, based on Genetic Algorithms, for the control of the 'security level' of a Cyber-Physical System (CPS). The security level is a quantity that has been studied in several industrial standards, among which we selected the Open Source Security Testing Methodology Manual (OSSTMM). The proposed optimization solution is validated on scenarios representative of real operations of a security evaluator, and numerical simulations report the performances obtained by the algorithm. © 2019 IEEE.

@CONFERENCE{Martirano2018,
author={Martirano, L. and Habib, E. and Giuseppi, A. and Di Giorgio, A.},
title={Nearly zero energy building model predictive control for efficient heating},
journal={2018 IEEE Industry Applications Society Annual Meeting, IAS 2018},
year={2018},
doi={10.1109/IAS.2018.8544632},
art_number={8544632},
abstract={Residential and non-residential buildings are responsible for approximately 40% of energy consumption and CO2 emissions in the EU. Considering that almost 75% of the building stock in EU is energy inefficient, the European energy policy promotes the improvement of the energy performance of existing buildings by introducing the innovative model of nearly zero energy building (nZEB). In the nZEB model, local energy sources (generation, storage and load management), building automation (BACS) and electronic monitoring of technical building systems (TBS) play a fundamental role. In electric systems, smart grids are a key feature of future energy scenarios, with the overarching goal of better aligning energy generation and demand. The challenge is the role of the users. The nZEB model with its “smart microgrid” can represent an effective driver according to the new policies of user's aggregation. In this framework demand side management (DSM) strategies can be implemented. The paper presents an innovative approach to use BACS present in nZEBS not only to increase the efficiency of TBS but also to operate an energy storage by heating systems for DSM strategies. © 2018 IEEE},
author_keywords={Building management systems;  Demand side management;  Energy storage;  Heat pumps;  Nearly zero energy building},
keywords={Demand side management;  Electric power transmission networks;  Energy efficiency;  Energy management systems;  Energy storage;  Energy utilization;  Housing;  Intelligent buildings;  Model predictive control;  Smart power grids;  Storage management, Building automation;  Building management system;  Demand Side Management (DSM);  Electronic monitoring;  European energy policy;  Heat pumps;  Innovative approaches;  Residential building, Zero energy buildings},
document_type={Conference Paper},
}

Residential and non-residential buildings are responsible for approximately 40% of energy consumption and CO2 emissions in the EU. Considering that almost 75% of the building stock in EU is energy inefficient, the European energy policy promotes the improvement of the energy performance of existing buildings by introducing the innovative model of nearly zero energy building (nZEB). In the nZEB model, local energy sources (generation, storage and load management), building automation (BACS) and electronic monitoring of technical building systems (TBS) play a fundamental role. In electric systems, smart grids are a key feature of future energy scenarios, with the overarching goal of better aligning energy generation and demand. The challenge is the role of the users. The nZEB model with its “smart microgrid” can represent an effective driver according to the new policies of user's aggregation. In this framework demand side management (DSM) strategies can be implemented. The paper presents an innovative approach to use BACS present in nZEBS not only to increase the efficiency of TBS but also to operate an energy storage by heating systems for DSM strategies. © 2018 IEEE

@CONFERENCE{Giuseppi2018210,
author={Giuseppi, A. and Germana, R. and Di Giorgio, A.},
title={Risk Adverse Virtual Power Plant Control in Unsecure Power Systems},
journal={MED 2018 - 26th Mediterranean Conference on Control and Automation},
year={2018},
pages={210-216},
doi={10.1109/MED.2018.8442768},
art_number={8442768},
abstract={This paper presents a control strategy for enabling a large scale Virtual Power Plant (VPP) constituted by a traditional power plant, distributed Energy Storage Systems (ESSs) and wind turbine driven Doubly Fed Induction Generators (DFIGs) to virtual slack bus functions in electricity transmission networks. The VPP in question is in charge of covering the network losses and a portion of the day ahead generation schedule of unsecured power plants, in presence of short term notifications about possible malicious/natural adverse events affecting them. The objective is pursued by integrating a dynamic optimal power flow problem in a realtime Model Predictive Control framework, and applying a second level of control aimed at keeping the dynamics of the real nonlinear plant subject to wind turbulence in line with the dynamics of the MPC model. Simulation results provide a proof of the proposed concept, showing as the joint coordination of storage devices and wind turbines can be part of the task of providing support actions to the network traditionally delivered by expensive and pollutant legacy power plants. © 2018 IEEE.},
keywords={Asynchronous generators;  Electric energy storage;  Electric load flow;  Electric machine control;  Electric power system control;  Electric power transmission;  Model predictive control;  Predictive control systems;  Virtual storage;  Wind turbines, Control strategies;  Distributed energy storage systems;  Doubly fed induction generators;  Dynamic optimal power flow;  Electricity transmission networks;  Generation schedules;  Virtual power plants;  Virtual power plants (VPP), Electric power transmission networks},
document_type={Conference Paper},
}

This paper presents a control strategy for enabling a large scale Virtual Power Plant (VPP) constituted by a traditional power plant, distributed Energy Storage Systems (ESSs) and wind turbine driven Doubly Fed Induction Generators (DFIGs) to virtual slack bus functions in electricity transmission networks. The VPP in question is in charge of covering the network losses and a portion of the day ahead generation schedule of unsecured power plants, in presence of short term notifications about possible malicious/natural adverse events affecting them. The objective is pursued by integrating a dynamic optimal power flow problem in a realtime Model Predictive Control framework, and applying a second level of control aimed at keeping the dynamics of the real nonlinear plant subject to wind turbulence in line with the dynamics of the MPC model. Simulation results provide a proof of the proposed concept, showing as the joint coordination of storage devices and wind turbines can be part of the task of providing support actions to the network traditionally delivered by expensive and pollutant legacy power plants. © 2018 IEEE.

@CONFERENCE{Suraci2018466,
author={Suraci, V. and Celsi, L.R. and Giuseppi, A. and Manfredi, G. and Di Giorgio, A.},
title={Distributed Wardrop Load Balancing in Multi-MTU SCADA Systems},
journal={MED 2018 - 26th Mediterranean Conference on Control and Automation},
year={2018},
pages={466-472},
doi={10.1109/MED.2018.8442485},
art_number={8442485},
abstract={This paper presents a distributed strategy for load balancing in a multi-MTU SCADA system, whose automatic control layer is such that its MTU Plane is modeled as a networked dynamical system. The proposed control law, under which the considered system is proven to converge to a Wardrop equilibrium, is also used for the purpose of equilibrium recovery in load distribution among MTUs after the occurrence of a possible MTU failure event induced by a cyber-physical attack (e.g., a Denial of Service attack). Numerical simulations with respect to realistic scenarios are reported to show the effectiveness of the proposed approach. © 2018 IEEE.},
keywords={Automation;  Denial-of-service attack;  Dynamical systems;  Networked control systems, Control laws;  Cyber physicals;  Distributed strategies;  Failure events;  Load distributions;  Networked dynamical systems;  Realistic scenario;  Wardrop equilibrium, SCADA systems},
document_type={Conference Paper},
}

This paper presents a distributed strategy for load balancing in a multi-MTU SCADA system, whose automatic control layer is such that its MTU Plane is modeled as a networked dynamical system. The proposed control law, under which the considered system is proven to converge to a Wardrop equilibrium, is also used for the purpose of equilibrium recovery in load distribution among MTUs after the occurrence of a possible MTU failure event induced by a cyber-physical attack (e.g., a Denial of Service attack). Numerical simulations with respect to realistic scenarios are reported to show the effectiveness of the proposed approach. © 2018 IEEE.

@CONFERENCE{Panfili2018460,
author={Panfili, M. and Giuseppi, A. and Fiaschetti, A. and Al-Jibreen, H.B. and Pietrabissa, A. and Delli Priscoli, F.},
title={A Game-Theoretical Approach to Cyber-Security of Critical Infrastructures Based on Multi-Agent Reinforcement Learning},
journal={MED 2018 - 26th Mediterranean Conference on Control and Automation},
year={2018},
pages={460-465},
doi={10.1109/MED.2018.8442695},
art_number={8442695},
abstract={This paper presents a control strategy for Cyber-Physical System defense developed in the framework of the European Project ATENA, that concerns Critical Infrastructure (CI) protection. The aim of the controller is to find the optimal security configuration, in terms of countermeasures to implement, in order to address the system vulnerabilities. The attack/defense problem is modeled as a multi-agent general sum game, where the aim of the defender is to prevent the most damage possible by finding an optimal trade-off between prevention actions and their costs. The problem is solved utilizing Reinforcement Learning and simulation results provide a proof of the proposed concept, showing how the defender of the protected CI is able to minimize the damage caused by his her opponents by finding the Nash equilibrium of the game in the zero-sum variant, and, in a more general scenario, by driving the attacker in the position where the damage she/he can cause to the infrastructure is lower than the cost it has to sustain to enforce her/his attack strategy. © 2018 IEEE.},
author_keywords={Composable Security;  Critical Infrastructure Protection;  Reinforcement Learning;  Stochastic Games;  Vulnerability Management},
keywords={Costs;  Economic and social effects;  Embedded systems;  Game theory;  Multi agent systems;  Public works;  Reinforcement learning;  Stochastic systems, Composable;  Control strategies;  Critical infrastructure protection;  Multi-agent reinforcement learning;  Stochastic game;  System vulnerability;  Theoretical approach;  Vulnerability management, Critical infrastructures},
document_type={Conference Paper},
}

This paper presents a control strategy for Cyber-Physical System defense developed in the framework of the European Project ATENA, that concerns Critical Infrastructure (CI) protection. The aim of the controller is to find the optimal security configuration, in terms of countermeasures to implement, in order to address the system vulnerabilities. The attack/defense problem is modeled as a multi-agent general sum game, where the aim of the defender is to prevent the most damage possible by finding an optimal trade-off between prevention actions and their costs. The problem is solved utilizing Reinforcement Learning and simulation results provide a proof of the proposed concept, showing how the defender of the protected CI is able to minimize the damage caused by his her opponents by finding the Nash equilibrium of the game in the zero-sum variant, and, in a more general scenario, by driving the attacker in the position where the damage she/he can cause to the infrastructure is lower than the cost it has to sustain to enforce her/his attack strategy. © 2018 IEEE.

@ARTICLE{Pietrabissa2018,
author={Pietrabissa, A. and Ricciardi Celsi, L. and Cimorelli, F. and Suraci, V. and Delli Priscoli, F. and Di Giorgio, A. and Giuseppi, A. and Monaco, S.},
title={Lyapunov-Based Design of a Distributed Wardrop Load-Balancing Algorithm With Application to Software-Defined Networking},
journal={IEEE Transactions on Control Systems Technology},
year={2018},
doi={10.1109/TCST.2018.2842044},
abstract={This paper presents an original discrete-time, distributed, noncooperative load-balancing algorithm, based on mean field game theory, which does not require explicit communications. The algorithm is proven to converge to an arbitrarily small neighborhood of a specific equilibrium among the loads of the providers, known as Wardrop equilibrium. Thanks to its characteristics, the algorithm is suitable for the software-defined networking (SDN) scenario, where service requests coming from the network nodes, i.e., the switches, are managed by the so-called SDN controllers, playing the role of providers. The proposed approach is aimed at dynamically balancing the requests of the switches among the SDN controllers to avoid congestion. This paper also suggests the adoption of SDN Proxies to improve the scalability of the overall SDN paradigm and presents an implementation of the algorithm in a proof-of-concept SDN scenario, which shows the effectiveness of the proposed solution with respect to the current approaches. IEEE},
author_keywords={Control systems;  Game theory;  Heuristic algorithms;  Load balancing;  Load management;  Lyapunov design;  Software;  Software algorithms;  software-defined networks (SDN);  Time factors;  Wardrop equilibrium.},
keywords={Application programs;  Game theory;  Resource allocation, Explicit communication;  Load balancing algorithms;  Lyapunov based design;  Lyapunov design;  Proof of concept;  Service requests;  Software defined networking (SDN);  Wardrop equilibrium, Software defined networking},
document_type={Article},
}

This paper presents an original discrete-time, distributed, noncooperative load-balancing algorithm, based on mean field game theory, which does not require explicit communications. The algorithm is proven to converge to an arbitrarily small neighborhood of a specific equilibrium among the loads of the providers, known as Wardrop equilibrium. Thanks to its characteristics, the algorithm is suitable for the software-defined networking (SDN) scenario, where service requests coming from the network nodes, i.e., the switches, are managed by the so-called SDN controllers, playing the role of providers. The proposed approach is aimed at dynamically balancing the requests of the switches among the SDN controllers to avoid congestion. This paper also suggests the adoption of SDN Proxies to improve the scalability of the overall SDN paradigm and presents an implementation of the algorithm in a proof-of-concept SDN scenario, which shows the effectiveness of the proposed solution with respect to the current approaches. IEEE

@ARTICLE{Adamsky201872,
author={Adamsky, F. and Aubigny, M. and Battisti, F. and Carli, M. and Cimorelli, F. and Cruz, T. and Di Giorgio, A. and Foglietta, C. and Galli, A. and Giuseppi, A. and Liberati, F. and Neri, A. and Panzieri, S. and Pascucci, F. and Proenca, J. and Pucci, P. and Rosa, L. and Soua, R.},
title={Integrated protection of industrial control systems from cyber-attacks: the ATENA approach},
journal={International Journal of Critical Infrastructure Protection},
year={2018},
volume={21},
pages={72-82},
doi={10.1016/j.ijcip.2018.04.004},
abstract={Industrial and Automation Control systems traditionally achieved security thanks to the use of proprietary protocols and isolation from the telecommunication networks. Nowadays, the advent of the Industrial Internet of Things poses new security challenges. In this paper, we first highlight the main security challenges that advocate for new risk assessment and security strategies. To this end, we propose a security framework and advanced tools to properly manage vulnerabilities, and to timely react to the threats. The proposed architecture fills the gap between computer science and control theoretic approaches. The physical layers connected to Industrial Control Systems are prone to disrupt when facing cyber-attacks. Considering the modules of the proposed architecture, we focus on the development of a practical framework to compare information about physical faults and cyber-attacks. This strategy is implemented in the ATENA architecture that has been designed as an innovative solution for the protection of critical assets. © 2018 Elsevier B.V.},
author_keywords={Critical infrastructures;  Cyber-physical attacks;  IACS;  Industrial and automation control systems;  Industrial IoT;  SCADA systems},
keywords={Computer crime;  Crime;  Critical infrastructures;  Internet of things;  Network architecture;  Network layers;  Risk assessment;  SCADA systems, Automation control system;  Control-theoretic approach;  Cyber physicals;  Industrial control systems;  Industrial IoT;  Integrated protection;  Proposed architectures;  Security, Network security},
document_type={Article},
}

Industrial and Automation Control systems traditionally achieved security thanks to the use of proprietary protocols and isolation from the telecommunication networks. Nowadays, the advent of the Industrial Internet of Things poses new security challenges. In this paper, we first highlight the main security challenges that advocate for new risk assessment and security strategies. To this end, we propose a security framework and advanced tools to properly manage vulnerabilities, and to timely react to the threats. The proposed architecture fills the gap between computer science and control theoretic approaches. The physical layers connected to Industrial Control Systems are prone to disrupt when facing cyber-attacks. Considering the modules of the proposed architecture, we focus on the development of a practical framework to compare information about physical faults and cyber-attacks. This strategy is implemented in the ATENA architecture that has been designed as an innovative solution for the protection of critical assets. © 2018 Elsevier B.V.

@ARTICLE{Liberati2017,
author={Liberati, F. and Giuseppi, A. and Pietrabissa, A. and Suraci, V. and Di Giorgio, A. and Trubian, M. and Dietrich, D. and Papadimitriou, P. and Delli Priscoli, F.},
title={Stochastic and exact methods for service mapping in virtualized network infrastructures},
journal={International Journal of Network Management},
year={2017},
volume={27},
number={6},
doi={10.1002/nem.1985},
art_number={e1985},
abstract={This paper presents a stochastic algorithm for virtual network service mapping in virtualized network infrastructures, based on reinforcement learning (RL). An exact mapping algorithm in line with the current state of the art and based on integer linear programming is proposed as well, and the performances of the two algorithms are compared. While most of the current works in literature report exact or heuristic mapping methods, the RL algorithm presented here is instead a stochastic one, based on Markov decision processes theory. The aim of the RL algorithm is to iteratively learn an efficient mapping policy, which could maximize the expected mapping reward in the long run. Based on the review of the state of the art, the paper presents a general model of the service mapping problem and the mathematical formulation of the 2 proposed strategies. The distinctive features of the 2 algorithms, their strengths, and possible drawbacks are discussed and validated by means of numeric simulations in a realistic emulated environment. Copyright © 2017 John Wiley & Sons, Ltd.},
keywords={Conformal mapping;  Decision theory;  Heuristic methods;  Integer programming;  Markov processes;  Reinforcement learning;  Stochastic systems, Integer Linear Programming;  Mapping algorithms;  Markov Decision Processes;  Mathematical formulation;  Network infrastructure;  Numeric simulation;  Stochastic algorithms;  Virtual network services, Iterative methods},
document_type={Article},
}

This paper presents a stochastic algorithm for virtual network service mapping in virtualized network infrastructures, based on reinforcement learning (RL). An exact mapping algorithm in line with the current state of the art and based on integer linear programming is proposed as well, and the performances of the two algorithms are compared. While most of the current works in literature report exact or heuristic mapping methods, the RL algorithm presented here is instead a stochastic one, based on Markov decision processes theory. The aim of the RL algorithm is to iteratively learn an efficient mapping policy, which could maximize the expected mapping reward in the long run. Based on the review of the state of the art, the paper presents a general model of the service mapping problem and the mathematical formulation of the 2 proposed strategies. The distinctive features of the 2 algorithms, their strengths, and possible drawbacks are discussed and validated by means of numeric simulations in a realistic emulated environment. Copyright © 2017 John Wiley & Sons, Ltd.

@CONFERENCE{Suraci2017761,
author={Suraci, V. and Celsi, L.R. and Giuseppi, A. and Di Giorgio, A.},
title={A distributed wardrop control algorithm for load balancing in smart grids},
journal={2017 25th Mediterranean Conference on Control and Automation, MED 2017},
year={2017},
pages={761-767},
doi={10.1109/MED.2017.7984210},
art_number={7984210},
abstract={This paper presents a distributed strategy for load balancing in a smart grid, modeling demand and supply as a networked dynamical system. The algorithm, which is characterized by point-to-point communications among agents implemented at the level of local energy management systems, is proven to converge to a Wardrop equilibrium. Numerical simulations of realistic scenarios are reported to show the effectiveness of the proposed approach. © 2017 IEEE.},
document_type={Conference Paper},
}

This paper presents a distributed strategy for load balancing in a smart grid, modeling demand and supply as a networked dynamical system. The algorithm, which is characterized by point-to-point communications among agents implemented at the level of local energy management systems, is proven to converge to a Wardrop equilibrium. Numerical simulations of realistic scenarios are reported to show the effectiveness of the proposed approach. © 2017 IEEE.

@CONFERENCE{DiGiorgio2017781,
author={Di Giorgio, A. and Giuseppi, A. and Liberati, F. and Pietrabissa, A.},
title={Controlled electricity distribution network black start with energy storage system support},
journal={2017 25th Mediterranean Conference on Control and Automation, MED 2017},
year={2017},
pages={781-786},
doi={10.1109/MED.2017.7984213},
art_number={7984213},
abstract={This paper presents an optimal procedure for controlled electricity black start for electricity distribution grids in reaction to adverse events or malicious cyber attacks resulting in the interruption of the main power supply from transmission network. The islanded operation is supposed to be guaranteed by an electric energy storage system, which covers the network imbalance between demand and supply from distributed generation during the sequential reconnection of medium voltage branches. The objective is to restore the service to the consumers while avoiding violation of technical constraints in terms of storage power flow and battery capacity. The discussion of simulation results assesses the effectiveness of the proposed approach in the context of a simplified network model and gives rise to relevant remarks and requirements for further developments. © 2017 IEEE.},
document_type={Conference Paper},
}

This paper presents an optimal procedure for controlled electricity black start for electricity distribution grids in reaction to adverse events or malicious cyber attacks resulting in the interruption of the main power supply from transmission network. The islanded operation is supposed to be guaranteed by an electric energy storage system, which covers the network imbalance between demand and supply from distributed generation during the sequential reconnection of medium voltage branches. The objective is to restore the service to the consumers while avoiding violation of technical constraints in terms of storage power flow and battery capacity. The discussion of simulation results assesses the effectiveness of the proposed approach in the context of a simplified network model and gives rise to relevant remarks and requirements for further developments. © 2017 IEEE.

@CONFERENCE{DiGiorgio2017986,
author={Di Giorgio, A. and Giuseppi, A. and Liberali, F. and Ornatelli, A. and Rabezzano, A. and Celsi, L.R.},
title={On the optimization of energy storage system placement for protecting power transmission grids against dynamic load altering attacks},
journal={2017 25th Mediterranean Conference on Control and Automation, MED 2017},
year={2017},
pages={986-992},
doi={10.1109/MED.2017.7984247},
art_number={7984247},
abstract={In this paper a power system protection scheme based on energy storage system placement against closed-loop dynamic load altering attacks is proposed. The protection design consists in formulating a non-convex optimization problem, subject to a Lyapunov stability constraint and solved using a two-step iterative procedure. Simulation results confirm the effectiveness of the approach and the potential relevance of using energy storage systems in support of primary frequency regulation services. © 2017 IEEE.},
keywords={Convex optimization;  Dynamic loads;  Energy storage;  Optimization, Closed loop dynamic;  Energy storage systems;  Lyapunov stability;  Nonconvex optimization;  Power system protection;  Power transmission grids;  Primary frequency regulation;  Protection designs, Electric power system protection},
document_type={Conference Paper},
}

In this paper a power system protection scheme based on energy storage system placement against closed-loop dynamic load altering attacks is proposed. The protection design consists in formulating a non-convex optimization problem, subject to a Lyapunov stability constraint and solved using a two-step iterative procedure. Simulation results confirm the effectiveness of the approach and the potential relevance of using energy storage systems in support of primary frequency regulation services. © 2017 IEEE.

@ARTICLE{Pietrabissa2017508,
author={Pietrabissa, A. and Priscoli, F.D. and Di Giorgio, A. and Giuseppi, A. and Panfili, M. and Suraci, V.},
title={An approximate dynamic programming approach to resource management in multi-cloud scenarios},
journal={International Journal of Control},
year={2017},
volume={90},
number={3},
pages={508-519},
doi={10.1080/00207179.2016.1185802},
abstract={The programmability and the virtualisation of network resources are crucial to deploy scalable Information and Communications Technology (ICT) services. The increasing demand of cloud services, mainly devoted to the storage and computing, requires a new functional element, the Cloud Management Broker (CMB), aimed at managing multiple cloud resources to meet the customers’ requirements and, simultaneously, to optimise their usage. This paper proposes a multi-cloud resource allocation algorithm that manages the resource requests with the aim of maximising the CMB revenue over time. The algorithm is based on Markov decision process modelling and relies on reinforcement learning techniques to find online an approximate solution. © 2016 Informa UK Limited, trading as Taylor & Francis Group.},
author_keywords={approximate dynamic programming;  Cloud networks;  Markov decision process;  reinforcement learning;  resource management},
keywords={Distributed computer systems;  Economics;  Markov processes;  Natural resources management;  Reinforcement learning;  Resource allocation, Approximate dynamic programming;  Approximate solution;  Cloud networks;  Information and communications technology;  Markov Decision Processes;  Reinforcement learning techniques;  Resource allocation algorithms;  Resource management, Dynamic programming},
document_type={Article},
}

The programmability and the virtualisation of network resources are crucial to deploy scalable Information and Communications Technology (ICT) services. The increasing demand of cloud services, mainly devoted to the storage and computing, requires a new functional element, the Cloud Management Broker (CMB), aimed at managing multiple cloud resources to meet the customers’ requirements and, simultaneously, to optimise their usage. This paper proposes a multi-cloud resource allocation algorithm that manages the resource requests with the aim of maximising the CMB revenue over time. The algorithm is based on Markov decision process modelling and relies on reinforcement learning techniques to find online an approximate solution. © 2016 Informa UK Limited, trading as Taylor & Francis Group.

@CONFERENCE{Riera2016243,
author={Riera, J.F. and Batalle, J. and Bonnet, J. and Dias, M. and McGrath, M. and Petralia, G. and Liberati, F. and Giuseppi, A. and Pietrabissa, A. and Ceselli, A. and Petrini, A. and Trubian, M. and Papadimitrou, P. and Dietrich, D. and Ramos, A. and Melian, J. and Xilouris, G. and Kourtis, A. and Kourtis, T. and Markakis, E.K.},
title={TeNOR: Steps towards an orchestration platform for multi-PoP NFV deployment},
journal={IEEE NETSOFT 2016 - 2016 IEEE NetSoft Conference and Workshops: Software-Defined Infrastructure for Networks, Clouds, IoT and Services},
year={2016},
pages={243-250},
doi={10.1109/NETSOFT.2016.7502419},
art_number={7502419},
abstract={Network Functions Visualization is focused on migrating traditional hardware-based network functions to software-based appliances running on standard high volume severs. There are a variety of challenges facing early adopters of Network Function Virtualizations; key among them are resource and service mapping, to support virtual network function orchestration. Service providers need efficient and effective mapping capabilities to optimally deploy network services. This paper describes TeNOR, a micro-service based network function virtualisation orchestrator capable of effectively addressing resource and network service mapping. The functional architecture and data models of TeNOR are described, as well as two proposed approaches to address the resource mapping problem. Key evaluation results are discussed and an assessment of the mapping approaches is performed in terms of the service acceptance ratio and scalability of the proposed approaches. © 2016 IEEE.},
author_keywords={Management and Orchestration;  Network service;  Service Mapping;  Virtual network function},
keywords={Mapping;  Virtual reality, Acceptance ratio;  Evaluation results;  Functional architecture;  Mapping capabilities;  Network functions;  Network services;  Resource mapping;  Virtual networks, Transfer functions},
document_type={Conference Paper},
}

Network Functions Visualization is focused on migrating traditional hardware-based network functions to software-based appliances running on standard high volume severs. There are a variety of challenges facing early adopters of Network Function Virtualizations; key among them are resource and service mapping, to support virtual network function orchestration. Service providers need efficient and effective mapping capabilities to optimally deploy network services. This paper describes TeNOR, a micro-service based network function virtualisation orchestrator capable of effectively addressing resource and network service mapping. The functional architecture and data models of TeNOR are described, as well as two proposed approaches to address the resource mapping problem. Key evaluation results are discussed and an assessment of the mapping approaches is performed in terms of the service acceptance ratio and scalability of the proposed approaches. © 2016 IEEE.

@CONFERENCE{Pietrabissa20159084,
author={Pietrabissa, A. and Battilotti, S. and Facchinei, F. and Giuseppi, A. and Oddi, G. and Panfili, M. and Suraci, V.},
title={Resource management in multi-cloud scenarios via reinforcement learning},
journal={Chinese Control Conference, CCC},
year={2015},
volume={2015-September},
pages={9084-9089},
doi={10.1109/ChiCC.2015.7261077},
art_number={7261077},
abstract={The concept of Virtualization of Network Resources, such as cloud storage and computing power, has become crucial to any business that needs dynamic IT resources. With virtualization, we refer to the migration of various tasks, usually performed by hardware infrastructures, to virtual IT resources. This approach allows resources to be rapidly deployed, scaled and dynamically reassigned. In the last few years, the demand of cloud resources has grown dramatically, and a new figure plays a key role: the Cloud Management Broker (CMB). The CMB purpose is to manage cloud resources to meet the user's requirements and, at the same time, to optimize their usage. This paper proposes two multi-cloud resource allocation algorithms that manage the resource requests with the aim of maximizing the CMB revenue over time. The algorithms, based on Reinforcement Learning techniques, are evaluated and compared by numerical simulations. © 2015 Technical Committee on Control Theory, Chinese Association of Automation.},
author_keywords={Cloud networks;  Markov Decision Process;  Reinforcement Learning;  Resource Management},
document_type={Conference Paper},
}

The concept of Virtualization of Network Resources, such as cloud storage and computing power, has become crucial to any business that needs dynamic IT resources. With virtualization, we refer to the migration of various tasks, usually performed by hardware infrastructures, to virtual IT resources. This approach allows resources to be rapidly deployed, scaled and dynamically reassigned. In the last few years, the demand of cloud resources has grown dramatically, and a new figure plays a key role: the Cloud Management Broker (CMB). The CMB purpose is to manage cloud resources to meet the user's requirements and, at the same time, to optimize their usage. This paper proposes two multi-cloud resource allocation algorithms that manage the resource requests with the aim of maximizing the CMB revenue over time. The algorithms, based on Reinforcement Learning techniques, are evaluated and compared by numerical simulations. © 2015 Technical Committee on Control Theory, Chinese Association of Automation.